Audio version of the article
If we want to mitigate climate change, we need to find cost-effective and sustainable ways to reduce industrial carbon dioxide (CO2) emissions. Unfortunately, most of the well-established carbon capture and storage (CCS) processes in industrial post-combustion sources have adverse cost, environmental toxicity, or durability issues. In this context, many researchers have focused on what could be our best choice for next generation CCS systems: CO2 adsorption using solid porous carbon materials.
A notable benefit of using porous carbons for CO2 sequestration is that they can be made from biomass residues such as agricultural residues, food waste, animal waste, and forest debris. This makes porous carbons from biomass waste (BWDPC) attractive not only because of their low cost, but also because they offer an alternative way of making sensible use of biomass waste. Although BWDPCs could definitely bring us closer to a circular economy, this field of study is relatively young, and no clear guidelines or consensus exist between scientists as to how the BWDPCs are to be synthesized or what properties and compositions of the materials they should pay attention to.
Could Artificial Intelligence (AI) help us with this puzzle? In a study recently published in Environmental Science and Technology, a collaborative research team from the University of Korea and the National University of Singapore used a machine learning-based approach that can guide the development of a porous future carbon synthesis strategies: Scientists observed that three key factors influence the adsorption properties of CO2 in BWDPCs: the elemental composition of the porous solid, its texture properties, and the adsorption parameters at which it operates, such as temperature and pressure. So far it is unclear how these core factors should be prioritized in the development of BWDPC.
To help settle this matter, the team first conducted a literature review and selected 76 publications describing both the synthesis and performance of various BWDPCs. After curation, these papers provided over 500 datapoints that were used to train and test three tree-based models. “The main purpose of our work was to elucidate how machine learning tools can be leveraged for predictive analytics and used to draw valuable insights into the process of CO2 adsorption using BWDPCs,” explains Professor Yong Sik Ok from Korea University, who led the study.
The input properties of the models were the three key factors, while the output was the level of CO2 adsorption. Although the models essentially become “black boxes” even after the training process, they can be used to make accurate predictions about the performance of BWDPCs based solely on the core factors considered, more importantly through the feature analysis, the research team determined the relative importance of each of the input features in order to make accurate predictions. In other words, they established which of the core factors is the most important to achieve high CO2 adsorption. The results show that the adsorption parameters contributed much more than the other two key factors to the models making correct predictions, underscoring the importance of optimizing the operating conditions in the first place. The textural properties of the BWDPCs, such as their pore size and surface came second, and its elemental composition came last.
Worth noting, the predictions of the models and the results of the feature importance analyses were backed by existing literature and our current understanding of the mechanisms behind the CO2 capture process. This cemented the real-world applicability of this data-driven strategy not only for BWDPCs, but for other types of materials, as Prof. Ok explains, “Our modeling approach is cross-deployable and can be used to investigate other types of porous carbons for CO2 adsorption, such as zeolites and metal−organic frameworks, and not just those derived from biomass waste.”
The team now plans to devise a synthesis strategy for BWDPC that will focus on optimizing the two most important key factors. Additionally, they will continue to add experimental data points to the database used in this study and make it open source for the benefit of the research community as well.